In this cross-sectional study we compared the abilities of lipoprotein(a) [Lp(a)], plasminogen activator inhibitor-1 (PAI-1), and tissue plasminogen activator (TPA) to discriminate between individuals with and without a history of stroke from among subjects in a metabolic ward. A total of 210 subjects (108 men and 102 women; mean age, 63.8 years; range, 31 to 86 years) provided plasma and DNA samples for the study. Of these, 51 men and 50 women had a history of ischemic stroke. The 109 subjects without a history of stroke were compared with those with such a history for major risk factors for ischemic events. Mean plasma TPA and PAI-1 levels significantly (P < .001) discriminated among subjects younger than 70 years with a history of stroke. The mean plasma Lp(a) level of stroke subjects (21.9 mg/dL) did not differ significantly from that of control subjects (15.2 mg/dL). However, among individuals < 70 years old, Lp(a) plasma levels > 50 mg/dL were more common among stroke patients (8 with versus 1 without, P < .01 by chi 2 test). A molecular variation in the 5' flanking region of the apo(a) gene that has been related to elevated Lp(a) plasma levels (G/A-914) was not strongly correlated with circulating levels of Lp(a), nor did Lp(a) levels correlate with a polymorphism of the apo(a) gene (G/A-21), which is strongly linked (P < .001) to the G/A-914 variation. In this setting, the relation between Lp(a) and cerebral ischemia appears to be limited to individuals below 70 years with elevated (> 50 mg/dL) plasma levels of the lipoprotein.